Wind and solar resource assessment and prediction using Artificial Neural Network and semi-empirical model: case study of the Colombian Caribbean region

Abstract

This work is focused on the importance of developing and promoting the use of wind and solar energy resources in the Colombian Caribbean coast. This region has a considerable interest for the development of solar technology due to the available climatic characteristics. Therefore, a detailed solarimetric analysis has been carried out in the department of San Andr es, Providencia and Santa Catalina, located in the Colombian Caribbean region, using a semi-empirical radiation model, based on the Bird & Hulstrom model, and the parameterizations of the M€ achler & Iqbal model, which allowed obtaining an average total irradiation value of 6.5 kWh/m 2 day. In addition, a statistical analysis of the wind resource was carried out based on meteorological data, which yielded an average multiannual wind speed of 3.4 m/s, and a maximum wind speed of 15.2 m/s during the month of October. The meteorological input data used for this analysis were provided by the Colombian Institute of Hydrology, Meteorology and Environmental Studies (IDEAM), in order to perform initial calculations and obtain a climatic profile of the areas with clear, medium and cloudy atmospheres throughout the year. Regarding the comparative study, the analysis was complemented with a prediction of solar radiation using Artificial Neural Networks (ANN), where irradiance could be predicted with a fairly good agreement, which was validated with a Root Mean Square Error (RMSE) of 0.87 using the temperature, the relative humidity, the pressure and the wind speed as the input data.

Prediction of gross calorific value from coal analysis using decision tree-based bagging and boosting techniques

“…Computer-based soft computing methods, such as support vector machine (SVM), particle swarm optimization (PSO), fuzzy logic (FL), fuzzy decision tree (FDT), artificial neural networks ( ), wavelet neural network (WNN), genetic algorithm ( ), adaptive neuro-fuzzy inference system (ANFIS), co-active neuro-fuzzy inference system (CANFIS), convolutional neural network (CNN), imperialist competitive algorithm ( ) and, recurrent neural network (RNN) have recently been advanced in research areas of scientific, engineering, technological, and industrial courses [ [16] , [17] , [18] , [19] , [20] , [21] , [22] , [23] ]. These state-of-the-art mathematical modeling tools can capture high dimensional complex data, recognize inherent highly complex links from input-output data, find optimum patterns, and forecast target parameters [ 24 ].…”

ANFIS and ANN models to predict heliostat tracking errors

“…Regardless of prior knowledge about the physical phenomenon and the nature of the relationships between the input/output variables, ANN and ANFIS networks can be used to interpret the behavior of complex nonlinear problems and, therefore, predict their future response. These two methods of prediction have been used in several fields [ [13] , [14] , [15] , [16] , [17] , [18] ]. To determine the most suitable technique for predicting data, ANN and ANFIS performances have been compared in many instances [ 19 , 20 ].…”

Implementation of the Hierarchical Analytical Process in the Selection of the Best Source of Renewable Energy in the Colombian Caribbean Region

“…In the work of (Moghadam and Lombardi, 2019) an economic, technological and environmental optimization model of energy generation projects is developed with the aim of minimizing greenhouse gases, economic energy costs and increasing energy efficiency; the uncertainty treatment was carried out using Monte Carlo simulation (Milanés-Hermosilla et al, 2021). These investigations, both nationally and internationally, help to set an important precedent for future research on energy planning in Colombia and serve as a starting point for our work focused on energy planning in the Colombian Caribbean region (Silvera et al 2021;Zanghelini et al, 2018).…”